Laser & Optoelectronics Progress, Volume. 58, Issue 3, 3300061(2021)

Identification of X-Ray Fluorescent Spectral Paper Ashes Based on Support Vector Machine Algorithm

Li Chunyu1, Liu Jinkun1, Jiang Hong1、*, Xu Lele1, and Man Ji2
Author Affiliations
  • 1School of Forensic Science, People''s Public Security University of China, Beijing 100038, China
  • 2Beijing Huayi Hongsheng Technology Co. Ltd., Beijing 100123, China
  • show less
    Figures & Tables(9)
    SVM system structure
    SVM training model
    SVM learning results for training set(accuracy is 98.8235%)
    SVM prediction results for test set(accuracy is 100%)
    • Table 1. Paper sample

      View table

      Table 1. Paper sample

      LabelMaterialDescription
      1Work paperYu Rongfeng’s Notebook
      2Work paperM&G scrap book
      3Work paperPPSUC’s scrap book
      4Work paperPPSUC’s note book
      5Wrap paperBASTO shoes box
      6Wrap paperGuirenniao shoes box
      7Rice paperCopybook depict paper
      8Copybook paperCopybook paper
      9Postcard paperPPSUC postcard
      10Postcard paperEnvelope
      11Ivory board paperLianhua Qingwen wrap box
      12Ivory board paperSITILON mask box
      13Chip board paperFranzzi wrap box
      14Chip board paperCROWN bread box
      15Ivory board paperBASTO shoes box
      16Thin plastic paperViTa tea box
      17Ivory board paperDHC box
      18Ivory board paperMAC box
      19Ivory board paperSock box
      20Ivory board paperAuberge box
      21Pictorial paperPPSUC keepsake box
      22Pictorial paperAlipay advertising brochure
      23Pictorial paperHoliland advertising brochure
      24Instructions paperAuberge instructions
      25Invoice paperPPSUC invoice
      26Double side paperFruit knife wrap box
      27Name card paperCartoon card
      28Express bill paperZTO express bill
      29Oil surface sticker paper

      White oil surface

      sticker paper

      30Oil surface sticker paper

      Yellow oil surface

      sticker paper

    • Table 2. Feature elements of chemical composition in paper

      View table

      Table 2. Feature elements of chemical composition in paper

      Chemical compositionFunctionFeature element
      CaCO3To increase impermeability and gloss of paperCa
      ZnSO4Whitening agentZn
      ClO2BleachCl
      HBrFungicides and inhibitorsBr
      TiO2Whitening agentTi
      PbPrinting inkPb
      FeSO4Compound salt and plant absorptionFe
      Na2SnO2Preservatives and fuel-assisted agentsSn
      CuSO4Better flocculation and de-impurityCu
      NiOCatalystNi
      CoO3Catalysts and dyesCo
    • Table 3. Proportion validation table of data in training set to data in test set

      View table

      Table 3. Proportion validation table of data in training set to data in test set

      ProportionMisjudged sample numberAccuracy /%
      5∶41682.20
      6∶31385.50
      7∶2692.20
      8∶1297.70
      17∶10100.00
    • Table 4. Sample element indicator mean table

      View table

      Table 4. Sample element indicator mean table

      SampleZnBrClCaTiFePbSnCoCuNi
      1#15183041626091224425411017
      20#13647035425694190394316016
    • Table 5. Accuracy comparison between original test set and new test set after adding mean value

      View table

      Table 5. Accuracy comparison between original test set and new test set after adding mean value

      ProportionAccuracy after adding mean/%Accuracy for original test set /%
      7∶28592.20
      8∶19097.70
      17∶1100100.00
    Tools

    Get Citation

    Copy Citation Text

    Li Chunyu, Liu Jinkun, Jiang Hong, Xu Lele, Man Ji. Identification of X-Ray Fluorescent Spectral Paper Ashes Based on Support Vector Machine Algorithm[J]. Laser & Optoelectronics Progress, 2021, 58(3): 3300061

    Download Citation

    EndNote(RIS)BibTexPlain Text
    Save article for my favorites
    Paper Information

    Category: Spectroscopy

    Received: May. 6, 2020

    Accepted: --

    Published Online: Mar. 12, 2021

    The Author Email: Hong Jiang (jiangh2001@163.com)

    DOI:10.3788/LOP202158.0330006

    Topics